EP0082224B1 - Semiconductor device passivated with glass material - Google Patents
Semiconductor device passivated with glass material Download PDFInfo
- Publication number
- EP0082224B1 EP0082224B1 EP19810306038 EP81306038A EP0082224B1 EP 0082224 B1 EP0082224 B1 EP 0082224B1 EP 19810306038 EP19810306038 EP 19810306038 EP 81306038 A EP81306038 A EP 81306038A EP 0082224 B1 EP0082224 B1 EP 0082224B1
- Authority
- EP
- European Patent Office
- Prior art keywords
- region
- moat
- semiconductor
- junction
- glass
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired
Links
Images
Classifications
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L23/00—Details of semiconductor or other solid state devices
- H01L23/28—Encapsulations, e.g. encapsulating layers, coatings, e.g. for protection
- H01L23/29—Encapsulations, e.g. encapsulating layers, coatings, e.g. for protection characterised by the material, e.g. carbon
- H01L23/291—Oxides or nitrides or carbides, e.g. ceramics, glass
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L23/00—Details of semiconductor or other solid state devices
- H01L23/28—Encapsulations, e.g. encapsulating layers, coatings, e.g. for protection
- H01L23/31—Encapsulations, e.g. encapsulating layers, coatings, e.g. for protection characterised by the arrangement or shape
- H01L23/3157—Partial encapsulation or coating
- H01L23/3178—Coating or filling in grooves made in the semiconductor body
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
- H01L2924/0001—Technical content checked by a classifier
- H01L2924/0002—Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00
Definitions
- the present invention relates to a semiconductor device in which an exposed surface of a semiconductor substrate, to which at least one pn junction is exposed, is passivated with glass material, and more particularly to a semiconductor device in which at least one pn junction is exposed in a moat provided in a semiconductor substrate, and the surface of the groove or moat including the exposed end of the pn junction is passivated with glass material.
- a surface part of a semiconductor substrate, to which a pn junction formed in the semiconductor substrate is exposed is protected with an insulating material, namely, a dielectric material.
- an insulating material namely, a dielectric material.
- harmful substances from the outside of the semiconductor device for example, ions of an alkali metal, water (moisture), or dust, may exert an undesirable influence upon the above surface part, to deteriorate the electric characteristics of the semiconductor device (for example, reverse blocking characteristics and/or leakage characteristics).
- Oxide or nitride of silicon, oxide of a metal, resin, and others have hitherto been used as the above-mentioned insulating material.
- Glass can form a thick film as compared with conventional insulating inorganic materials, is not readily subjected to influence from the outside, and can form a dense film. Also, a glass film has a high permeation-preventing effect for gas and liquid, as compared with an organic material film.
- Glass materials suitable for use in passivation for semiconductor devices include, for example, zinc borosilicate glass which contains ZnO as one of the main components, lead borosilicate glass and lead aluminosilicate glass both of which contain PbO and Si0 2 as two of the main components.
- a semiconductor device using glass as passivation material is preferably provided with a groove or moat in one major surface of a semiconductor substrate, and a terminal end of a pn junction to be passivated with glass is exposed to the surface of the groove or moat. Glass is applied in the groove or moat and is then fired to passivate the exposed end of the pn junction with glass.
- Such a structure advantageously enhances the blocking voltage of the device, and moreover makes it easy to fabricate the semiconductor device.
- a semiconductor device having such a structure is hereinafter referred to as a moat type semiconductor device.
- this device has a channel stopper region in the substrate adjacent and outside the moat, the channel stopper region being of the same conductivity type as the first one of the semiconductor regions but higher in impurity concentration than that region.
- this device corresponds to the pre-characterising part of claim 1.
- JP-A-54-14677 does not discuss the characteristics of glass used for passivation.
- a moat type semiconductor device provided with a channel stopper region may not provide a sufficiently high blocking voltage.
- electronic apparatuses have been in the form of solid state circuits, and therefore a semiconductor device having a blocking voltage of about 1200 V or more is desirable so as to control or rectify a current under a high applied voltage in the electronic apparatuses.
- a horizontal deflection circuit incorporated in a television receiver set requires a transistor having a blocking voltage higher than 1500 V, and a converter for 220 V line requires a gate turn off thyristor having a blocking voltage of at least 1200 V. It is difficult reproducibly to obtain a high blocking voltage of 1200 V or more with conventional moat type semiconductor devices.
- the present invention proposes that the fired glass has electric charges of a polarity for inducing charges of a polarity opposite to the conductivity type of the first one of the semiconductor regions in a part of that region in contact with the moat.
- the surface of the glass has negative charge, which induces positive charge in the part of the substrate contiguous to the groove or moat.
- the glass preferably has a negative surface charge density within a range from -3x10"cm- 2 to -9x10"cm- z , where the minus sign indicates the negative polarity of the surface charge.
- the blocking voltage of the above-mentioned pn junction depends upon the angle between a slope of the inner side surface of the moat and a pn junction plane terminating at the slope substended by a more heavily doped one of the pn junction-forming regions, and upon a surface charge density of glass which fills the moat.
- the blocking voltage of the pn junction can be improved by coating the moat with glass having negative surface charge in the case where an n-type semiconductor region is contiguous to the moat.
- the characteristics of the glass-semiconductor interface is also influenced to some extent by the orientation of semiconductor surface and that the blocking voltage is little influenced by the thickness of the glass layer in case when the thickness is not less than about 15 pm.
- the depth of the channel stopper region from the major surface of the substrate is greater than the depth of the pn junction (J) from that major surface.
- J pn junction
- Figs. 1A and 1B show a gate turn off thyristor (hereinafter referred to as a "GTO") according to a first embodiment of the present invention.
- a silicon substrate 1 has a pair of major surfaces 101 and 102, and a four-layer laminate structure made up of an n-type emitter, a p-type base region 12, an n-type base region 11 and a p-type emitter region 15 is formed between the major surfaces 101 and 102.
- the n-type base region 11 and the p-type base region 12 form the two semiconductor regions of the present invention.
- the n-emitter 13 is divided into a plurality of strip regions 13.
- Each of the n-emitter regions 13 thus formed is exposed at one major surface 101, and a cathode electrode 31 is provided on the exposed surface of each n-emitter region.
- the p-base region 12 is exposed to the major surface 101 so as to surround the n-emitter regions 13, and a gate electrode 32 is provided on the exposed surface of the p-base region 12.
- the cathode electrode 31 includes a common connecting region 310 for connecting in common divided electrode portions in ohmic contact with the respective n-emitter regions.
- the common connecting region 310 is provided above the exposed surface of the p-base region 12, with an Si0 2 film (not shown) interposed between the common connecting region 310 and the exposed surface of the p-base region 12 to insulate the common connecting region 310 from the base region 12.
- the p-emitter region 15 is provided in part of the other major surface 102 in approximate registry with the n-emitter regions. That is, the p-emitter region 15 is divided into a plurality of regions similar to and to oppose the divided n-emitter regions 13.
- a first n +- type anode short region 161 penetrates each of the p-emitter regions thus formed; at central part thereof.
- a second n +- type anode short region 162 is exposed at part of the major surface 102 other than respective exposed surfaces of the p-emitter regions and respective exposed surfaces of the first n + -type anode short regions 161.
- An anode electrode 33 is formed over all the major surfaces 102.
- the first anode short regions 161 and the second anode short region 162 form a short circuit between the n-base region 11 and the anode electrode 33. Such a shorting structure is effective in improving the turn-off characteristic.
- a moat 51 in the form of a closed loop is cut in the major surface 101 along the periphery thereof.
- the moat 51 is cut to a depth such that the bottom of the moat is within the high resistivity n-base region 11. Therefore, the pn junction formed between the p-base region 12 and the high resistivity n-base region 11 terminates at an inclined surface of the moat 51.
- An n t- type channel stopper region 14, which has a greater thickness than the p-base region 12, is formed as a closed loop outside the moat 51. Glass material 21 having characteristics which will be explained later, is fired in the moat 51. Part of the major surface 101 other than the part provided with the electrodes is coated with an Si0 2 film 41 for the purpose of passivation.
- the field mitigating means in the present embodiment are made up of the glass material 21 and the n T- type channel stopper region 14 which has a greater thickness than the p-base region 12 and is highly doped.
- Such a GTO can be fabricated, for example, by impurity diffusion techniques, photolithography and selective etching techniques, using an n-type (111) silicon plate having a resistivity of about 6-Qcm as the starting material.
- Gallium is diffused into the high resistivity n-type silicon plate from a pair of major surfaces thereof to form the P-base region 12 and p-emitter region 15.
- phosphorus is selectively diffused from the major surfaces into those areas corresponding to regions 14, 161 and 162. At this time, the phosphorus concentration in the diffused regions is greater than the gallium concentration in these regions, and the diffusion depth of the phosphorus is made greater than that of the gallium.
- the channel stopper region 14 is formed on the side of the major surface 101, and the first anode short regions 161 and second anode short region 162 are formed on the side of the major surface 102. Then, phosphorus is diffused into those areas corresponding to the n-emitter regions 13 so that the diffusion depth of the phosphorus is smaller than the depth of the p-base region 12, to form the n-emitter regions 13.
- the moat 51 can be made using selective etching techniques, and the electrodes 31, 32 and 33 can be formed by depositing a metal such as chromium, nickel, silver or aluminium on predetermined areas of the silicon substrate by e.g. vacuum evaporation techniques or the like.
- the SiO 2 film 41 can be formed by thermally oxidizing the exposed surface of the silicon substrate.
- Each n-emitter region 13 has a thickness of about 15 ⁇ m, and the surface of diffusion has an impurity concentration of about 10 20 atoms/cm 3 or more.
- the p-base region 12 has a thickness of about 30 ⁇ m, and the surface of diffusion has an impurity concentration of about 10 18 atoms/cm 3 .
- the n + -type channel stopper region 14 has a thickness of about 70 ⁇ m and a width of about 50 ⁇ m or more, and the surface of diffusion has an impurity concentration of about 10 20 atoms/cm 3 or more.
- the n-base region 11 has a thickness of about 180 pm and a resistivity of about 60 Ocm.
- the moat 51 has a width of about 300 ⁇ m and a depth of about 90 um, and therefore the depth of the moat is greater than the thickness of the channel stopper region 14 and also than that of the p-base region 12.
- the glass 21 is filled in the moat 51 provided in the above-mentioned GTO.
- the glass 21 can be formed by suspending glass powder in a liquid, adhering the powder to the surface of the moat 51 by electrophoresis, and then firing the powder.
- the glass 21 obtained after firing has a thickness of 20 to 30 ⁇ m.
- the glass material 21 used was zinc borosilicate glass.
- the zinc borosilicate glass preferably contains more than about 50 percent ZnO by weight.
- the amount of surface charge on the glass 21 can be varied by changing the composition of the glass 21, the firing temperature, firing time, firing atmosphere and the cooling rate after firing of the glass 21.
- Changes in the forward blocking voltage of the above-mentioned GTO namely, changes in the blocking voltage of the pn junction formed between the p-base region 12 and the n-base region 11
- the results of measurements are shown in Fig. 2.
- the amount of charge per unit area at the surface of the glass 21 which is in contact with the inner wall of the moat 51, is expressed by N FB (cm -2 ).
- the blocking voltage is highest when the surface charge density N FB is within a range from -3x10 11 /cm 2 to -9x10"/ cm 2 , that is, when the negative surface charge per unit area lies within a range from 3x10 11 /cm 2 to 9x10"/cm 2 .
- the high blocking voltage is not affected. In other words, a reproducible, high blocking voltage is obtained, even if the surface charge density N FB fluctuates.
- the glass 21 has a surface charge density N FB of -3x 10 11 /cm 2 to -9x10 11 /cm 2 .
- the zinc borosilicate glass had, for example, 65 percent ZnO, 21 percent B 2 0 3 , 8.5 percent Si0 2 , 4 percent PbO, 1 percent SnO and 0.5 percent Sb 2 0 3 by weight.
- the depth of the n *-type channel stopper region 14 is larger than that of the p-base region 12.
- Such a structure has the following advantages. Firstly, it is not necessary that a p-type impurity is selectively diffused into the silicon substrate 1 from the major surface 101 to form the p-base region 12.
- the channel stopper region 14 when the p-type impurity is diffused to a predetermined depth from the whole of the major surface 101 and then an n-type impurity is selectively diffused into the channel stopper region 14 in such a manner that the concentration of the n-type impurity is higher than that of the p-type impurity and the diffusion depth of the n-type impurity is greater than that of the p-type impurity, the channel stopper region shown in Fig. 1B is formed. Therefore, the fabricating process can be simplified. Such a fabricating method is particularly advantageous when gallium and aluminium are diffused as the p-type impurity, since it is difficult to obtain an appropriate mask for selective diffusion of gallium and aluminium.
- the region 14 were smaller in thickness than the region 12. In this case, it would be required to selectively form the region 12 in that part of the major surface 101 which is surrounded by the moat 51. If, in the above selective formation process, pin holes are accidentally generated in a cover for masking a portion of the major surface 101 outside the moat 51 from a diffusion source, p-type regions which are greater in depth than the region 14, are formed in the surface portion outside the moat 51, and the function of the channel stopper region is deteriorated. In other words, the effective width of the channel stopper region is reduced. Further, the yield of high breakdown voltage semiconductor devices is lowered. According to the present embodiment, all of these drawbacks may be eliminated.
- the gradient of impurity concentration at the boundary between the n +- type channel stopper region 14 and the n-base region 11 becomes gentle, since the thickness of the region 14 is large. Therefore, when the pn junction exposed at the moat 51 is reversely biased, the depletion layer extends from the pn junction into the n-base region 11 and, even after it has reached the channel stopper region 14, it can be further extended gradually in the channel stopper region 14 when the reverse bias voltage is increased. Thus, even when a high reverse bias voltage is applied, concentration of the electric field is mitigated, and therefore it is possible to provide a high breakdown voltage.
- the width of the channel stopper region 14 (that is, the width of the region 14 at the boundary between the region 14 and the n-base region 11 in the direction perpendicular to the lengthwise direction of the moat 51) is preferably made larger than 50 pm, in order effectively to extend the depletion layer and to make the breakdown voltage high.
- the depth of the moat 51 is preferably larger than respective thicknesses of the p-base region 12 and channel stopper region 14.
- the breakdown voltage varies widely. This is because the electric field generated at one end of the pn junction J, when the pn junction J is reverse biased, depends upon the length of that portion of the p-base region which is located beneath the bottom of the moat 51 and upon the depth of the moat 51 in a critical manner, and cannot be easily controlled.
- the difference in depth between the bottom of the moat 51 and the bottom of the region 12 is preferably made larger than 10 pm to reduce variations in the breakdown voltage.
- the reason for this is as follows. Compensation of impurity concentration takes place in the proximity of a pn junction, since two regions of different conductivity types are contiguous to each other at the pn junction, so that compensated regions have higher resistivity. Thus, the depletion layer can be readily extended in such regions.
- the pn junction is located in the proximity of the bottom of the moat 51, the depletion layer can be easily extended to reach the bottom of the moat 51 and further extension of the depletion layer is restricted. Therefore, the blocking voltage can be lowered at a portion beneath the moat 51, and the leakage current can be increased.
- the channel stopper region is formed by diffusion, it is desirable to make time required for the formation of the channel stopper region as short as possible. Accordingly, it is preferable to make the thickness of the channel stopper region smaller than the depth of the moat 51.
- the second embodiment relates to a reverse blocking thyristor.
- the reverse blocking thyristor must be able to block both forward and reverse voltages applied between anode and cathode electodes. That is, a high blocking voltage is desired in both forward and reverse directions.
- two moats 511 and 512 are provided in one major surface 101, and the depth and width of each of the moats 511 and 512 are nearly equal to the depth and width of the moat 51 shown in Fig. 1B.
- the distance between the moats 511 and 512 is about 70 pm.
- Cathode and gate electrodes 31 and 32 are formed on that inner part of the major surface 101 which is surrounded by the inner moat 511.
- the gate electrode 32 is near one corner of the above-mentioned inner part, and has substantially a rectangular form.
- the cathode electrode 31 is insulated from the gate electrode 32, and is formed on the rest of the above inner part. Such an electrode arrangement is called a corner gate type.
- Fig. 38 shows part of a cross section taken along the line IIIB ⁇ IIIB' in Fig. 3A.
- an n-emitter region 13, a p-base region 12, an n-base region 11 and a p-emitter region 15 are stacked in the order described, between the cathode electrode 31 and the anode electrode 32 formed on the other major surface 102.
- the n-base 11 region and the p-base region 12 form the two semiconductor regions of the present invention.
- the n-emitter region 13 is not divided into a plurality of regions, and the p-emitter region 15 has no shorting structure and is uniform in structure.
- a pn junction J, between the n-base region 11 and the p-emitter region 15 terminates at the outer moat 512 due to the presence of a p +- type region 151 (communicating with the p-emitter region 15) which connects peripheral parts of the major surfaces 101 and 102.
- a pn junction J 2 between the p-base region 12 and the n-base region 11 terminates at the inner inclined surface of the inner moat 511, and the pn junction J 2 and the inner inclined surface of the moat 511 form a negative bevel structure.
- An n +- type channel stopper region 14 is formed between the moats 511 and 512, exposed at the major surface 101, and is in the form of a closed loop.
- field mitigating means include the moats 511 and 512 filled with negatively charged glass material 21 and the channel stopper region 14.
- the semiconductor element shown in Figs. 3A and 3B can be fabricated by e.g. impurity diffusion techniques, photolithography, and selective etching techniques, as for the embodiment shown in Figs. 1A and 1B. Further, the same materials as used in the embodiment shown in Fig. 1A and 1B can be used.
- the thickness of the region 14 is greater than that of the n-emitter region 13 as in the embodiment shown in Figs. 1A and 1B.
- the p +- type region 151 may be formed e.g. by diffusing a p-type impurity from respective peripheral portions of the major surfaces 101 and 102 into the semiconductor substrate to a depth corresponding to more than one half of the thickness of the semiconductor substrate so that two diffused regions meet each other.
- a depletion layer extends from the pn junction J 2 mainly into the n-base region 11. Since the glass material 21 has negative charge, the depletion layer is more easily extended along the moat 511 rather than into the bulk of the semiconductor substrate in the area away from the moat. Therefore, electric field in the region around the moat is mitigated.
- the voltage applied between the anode and cathode electrodes is increased, the end of the depletion layer will reach the channel stopper region 14.
- the rate of extension of the depletion layer will be largely slowed down in the channel stopper region 14.
- the depletion layer is prevented from being further extended outward and from being exposed to the periphery of the semiconductor substrate. Accordingly, it is possible to maintain a high blocking voltage.
- the rate of extension of the depletion layer in the portion of the region 11 along the moat is determined mainly by the amount of charge kept by the glass material 21.
- a surface charge density N FB of the glass material 21 is preferably in a range from -3x 10"cm- 2 to -9x10 11 cm -2 to provide a breakdown voltage of about 1.6 kV, which has not been obtained heretofore.
- the thyristor takes a reverse blocking state.
- a depletion layer extends from the pn junction J, mainly into the n-base region 11.
- the rate of extension of the depletion layer becomes slow in the channel stopper region 14 in accordance with the same mechanism as in the forward blocking state. Accordingly even when a higher reverse voltage is applied, the depletion layer is not extended along the moat 511 and does not reach the p-base region 12, and the blocking state is not destroyed.
- the pn junction J 1 and the outer inclined surface of the moat 512 form a positive bevel structure
- the pn junction J 2 and the inner inclined surface of the moat 511 form a negative bevel structure
- the blocking ability in the forward blocking state is good but the blocking ability in the reverse blocking state is not so good.
- the blocking ability in the reverse blocking state is good but the blocking ability in the forward blocking state is not so good.
- the blocking function is good in both the forward and reverse blocking states, since the channel stopper region is formed and the glass 21 has negative charge.
- a semiconductor substrate 1 has a disc shape.
- Each major surface 101 and 102 has a circular shape, and the diameter of the major surface 101 is smaller than that of the major surface 102.
- the side surface 103 connecting the major surfaces 101 and 102 is not at right angles with the major surfaces 101 and 102, but is inclined to the major surfaces 101 and 102.
- a gate electrode 32 is formed on a central part of the major surface 101, and a cathode electrode 31 is formed on the major surface 101 so that it is insulated from and surrounds the gate electrode 32.
- a moat 51 in the form of a closed loop is formed around the cathode electrode 31.
- An anode electrode 33 is formed on the whole of the other major surface 102.
- the above-mentioned third embodiment is a reverse blocking thyristor having the same function as the reverse blocking thyristor shown in Figs. 3A and 3B.
- the thyristor of this embodiment as shown in Fig. 4A is called the reverse blocking thyristor of centre gate type, since the gate elctrode is arranged at a central portion of one major surface.
- Fig. 4B shows a part of a cross section taken along a diameter of the circular plate in Fig. 4A.
- an n-emitter region 13, a p-base region 12, an n-base region 11 and a p-emitter region 15 are stacked beween the cathode electrode 31 and the anode electrode 33, as in Fig. 3B.
- the n-base region 11 and the p-base region 12 form the two semiconductor regions of the present invention.
- the p-emitter region 15 is uniform and terminates at the side surface 103 of the semiconductor substrate 1.
- the p-emitter region 15 has a higher impurity concentration than the n-base region 11.
- the side surface 103 is tapered toward the major surface 101 such that the area of a plane parallel to the major surfaces 101 and 102 is diminished in a direction from the major surface 102 to 101. Therefore, the pn junction J 1 which is formed between the p-emitter region 15 and the n-base region 11, and the side surface 103 form a positive bevel structure. Such a positive bevel structure acts to increase the blocking voltage of the pn junction J" as described in US ⁇ A ⁇ 3491272.
- the pn junction J 2 between the p-base region 12 and the n-base region 11 terminates at the inner, inclined surface of the moat 51.
- the p-base region 12 has a higher impurity concentration than the n-base region 11.
- the inner inclined surface of the moat 51 and the pn junction J2 form - a negative bevel structure.
- the depth of the moat 51 is greater than the depth of the pn junction J 2 .
- a pn junction J 3 between the n-emitter region 13 and the p-base region 12, and the pn junction J 2 terminate at the inner inclined surface of the moat 51.
- the channel stopper region 14 is thicker than the n-emitter region 13. Glass material 21 which has an appropriate composition is applied in the moat 51 and fired to have negative charge.
- field mitigating means include the moat 51, the glass 21 in the moat and the positive bevel structure formed by the pn junction J, and the side surface 103.
- the semiconductor element shown in Figs. 4A and 4B can be fabricated in similar manner as the second embodiment shown in Figs. 3A and 3B, and therefore a detailed explanation is omitted. Further, the same materials as used in the second embodiment can be used in the third embodiment.
- a high blocking voltage is obtained on the basis of the same mechanism as in the embodiments shown in Figs. 1A and 3A.
- This glass essentially consists of three main components ZnO, B 2 0 3 and Si0 2 . Similar glass material is disclosed in DE-A-2851479.
- a region A indicates a composition range in which transparent glass was formed
- a region B indicates a composition range in which any mixed material was devitrified or divided into a plurality of phases and therefore transparent glass was not obtained.
- a glass material having a composition in the vitrifiable range shown in Fig. 5A was fused in a platinum crucible, and then quenched in the pure water to obtain a lump of glass, which was finely pulverized by a ball mill to obtain a glass powder for passivation.
- Various glass powders having different compositions within the vitrifiable range were prepared. Each of these glass powder materials was fired on a silicon substrate at a temperature in a range from 630°C to 780°C, and the surface charge density N FB was measured. In such a temperature range, glass is readily softened and is not crystallised.
- Fig. 6A shows equi-N FB curves which were obtained when the glass powder materials were fired at 630°C
- FIG. 6B shows equi-N FB curves for a firing temperature of 780°C.
- the value of N FB varies greatly with the composition, namely, the mixing ratio of the three components, and moves to the negative side as the firing temperature is higher.
- a composition range of glass and a range of firing temperature which are preferably applied to the present invention, are known from Figs. 6A and 6B.
- the coefficient of thermal expansion of the glass is as close as possible to that of the semiconductor material.
- moats as described in the above-mentioned embodiments are distributed almost all over one major surface of the wafer with glass filled therein.
- the semiconductor wafer may be warped due to the different thermal expansion (contraction) during cooling from the firing temperature to the room temperature.
- a warp may constitute an obstacle to the accurate printing of a predetermined photoresist pattern on the wafer. Further, when the warp of the wafer is large, there is a risk of the wafer and glass being cleaved and/or cracked. In the case of silicon, to avoid such difficulties, it is preferable to set the coefficient of thermal expansion of glass after firing within a range from 25x10 -7 /°C to 50xlO-7PC.
- the glass containing three main components ZnO, B 2 0 3 and Si0 2 as described above has the thermal expansion characteristics discussed below.
- Several glass powder materials having various compositions within the vitrifiable range shown in Fig. 5 were each fired to a thickness of 10 pm on a silicon wafer of 200 to 500 pm thick at a temperature within a range from 630 to 780°C, and the relation between the warp of silicon wafer and the difference in coefficient of thermal expansion between glass and silicon was studied. As a result of the study, it was found that the coefficient of thermal expansion of glass varied greatly with the composition of glass and the firing temperature.
- Fig. 7A shows equi-thermal-expansion curves in the case where glass powder materials were fired at 630°C
- FIG. 7B shows equi-thermal-expansion curves for a firing temperature of 780°C.
- glass containing a large amount of ZnO has a smaller coefficient of thermal expansion, and the equi-thermal-expansion curves move to a composition region containing a relatively smaller amount of ZnO, as the firing temperature is higher.
- the warp of silicon wafer was small and it was possible to print precisely a photoresist pattern for photo-etching. Moreover, the glass was not cracked.
- a composition range of glass and a range of firing temperature, which are preferable in the present invention, may be determined from Figs. 7A and 7B. Since the above-mentioned glass contains only three main components ZnO, B 2 0 3 and Si0 2 , variations in characteristic of glass are limited to a relatively small range. Therefore, when the above glass is used, uniform, reproducible semiconductor elements may be obtained.
- the present invention is not limited to these embodiments but is applicable to various semiconductor devices.
- a transistor structure is obtained.
- Such a structure is also within the scope of the present invention.
- the side surface 103 of the semiconductor substrate 1 may be coated with a passivation material such as glass or resin.
- the materials, fabricating methods and dimensions used in the embodiments may be appropriately altered.
- the conductivity type of each semiconductor region may be reversed. In this case, glass having positive charges is employed, and a preferable range of surface charge density is +3x10"cm- 2 to +9x10 11 cm -2.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Power Engineering (AREA)
- Chemical & Material Sciences (AREA)
- Ceramic Engineering (AREA)
- Thyristors (AREA)
Description
- The present invention relates to a semiconductor device in which an exposed surface of a semiconductor substrate, to which at least one pn junction is exposed, is passivated with glass material, and more particularly to a semiconductor device in which at least one pn junction is exposed in a moat provided in a semiconductor substrate, and the surface of the groove or moat including the exposed end of the pn junction is passivated with glass material.
- In semiconductor devices for rectifying or controlling a current flow such as a diode, a transistor or a thyristor, a surface part of a semiconductor substrate, to which a pn junction formed in the semiconductor substrate is exposed, is protected with an insulating material, namely, a dielectric material. Otherwise, harmful substances from the outside of the semiconductor device, for example, ions of an alkali metal, water (moisture), or dust, may exert an undesirable influence upon the above surface part, to deteriorate the electric characteristics of the semiconductor device (for example, reverse blocking characteristics and/or leakage characteristics).
- Oxide or nitride of silicon, oxide of a metal, resin, and others have hitherto been used as the above-mentioned insulating material.
- In recent years, attention has been attracted to preparing glass in the powder state and firing it on a semiconductor substrate for the same purpose as mentioned above. Glass can form a thick film as compared with conventional insulating inorganic materials, is not readily subjected to influence from the outside, and can form a dense film. Also, a glass film has a high permeation-preventing effect for gas and liquid, as compared with an organic material film.
- Glass materials suitable for use in passivation for semiconductor devices include, for example, zinc borosilicate glass which contains ZnO as one of the main components, lead borosilicate glass and lead aluminosilicate glass both of which contain PbO and Si02 as two of the main components.
- A semiconductor device using glass as passivation material is preferably provided with a groove or moat in one major surface of a semiconductor substrate, and a terminal end of a pn junction to be passivated with glass is exposed to the surface of the groove or moat. Glass is applied in the groove or moat and is then fired to passivate the exposed end of the pn junction with glass. Such a structure advantageously enhances the blocking voltage of the device, and moreover makes it easy to fabricate the semiconductor device. A semiconductor device having such a structure is hereinafter referred to as a moat type semiconductor device.
- An example of a moat type semiconductor device using glass as the passivation material is disclosed in JP-A-54^14677. This document shows a semiconductor device having:
- two semiconductor regions of opposite conductivity type in a semiconductor substrate, the two semiconductor regions defining a pn junction (J);
- a moat in the form of a closed loop in at least one major surface of the substrate such that the pn junction (J) terminates at an inner inclined surface of the moat, a first one of the semiconductor regions extending below the moat;
- fired glass in the moat for passivating an exposed end portion of the pn junction (J); and
- at least two electrodes on major surfaces of the substrate for permitting a reverse bias voltage to be applied to the pn junction (J).
- Furthermore this device has a channel stopper region in the substrate adjacent and outside the moat, the channel stopper region being of the same conductivity type as the first one of the semiconductor regions but higher in impurity concentration than that region. Thus this device corresponds to the pre-characterising part of
claim 1. - With this structure, even when the characteristic of an interface between glass and the semiconductor substrate varies, any adverse effect due to such a variation can be prevented by the channel stopper region, and therefore the high blocking voltage characteristic and high reliability is obtained. However, JP-A-54-14677 does not discuss the characteristics of glass used for passivation.
- However, detailed experiments made by the present inventors have revealed that even a moat type semiconductor device provided with a channel stopper region may not provide a sufficiently high blocking voltage. In recent years electronic apparatuses have been in the form of solid state circuits, and therefore a semiconductor device having a blocking voltage of about 1200 V or more is desirable so as to control or rectify a current under a high applied voltage in the electronic apparatuses. For example, a horizontal deflection circuit incorporated in a television receiver set requires a transistor having a blocking voltage higher than 1500 V, and a converter for 220 V line requires a gate turn off thyristor having a blocking voltage of at least 1200 V. It is difficult reproducibly to obtain a high blocking voltage of 1200 V or more with conventional moat type semiconductor devices.
- Therefore the present invention proposes that the fired glass has electric charges of a polarity for inducing charges of a polarity opposite to the conductivity type of the first one of the semiconductor regions in a part of that region in contact with the moat.
- Thus, when the semiconductor region which is contiguous to the bottom of the groove or moat, is an n-type region, the surface of the glass has negative charge, which induces positive charge in the part of the substrate contiguous to the groove or moat. The glass preferably has a negative surface charge density within a range from -3x10"cm-2 to -9x10"cm-z, where the minus sign indicates the negative polarity of the surface charge.
- Experiments made by the present inventors have revealed that, in moat type semiconductor devices each provided with a highly-doped channel stopper region, the blocking voltage of the above-mentioned pn junction depends upon the angle between a slope of the inner side surface of the moat and a pn junction plane terminating at the slope substended by a more heavily doped one of the pn junction-forming regions, and upon a surface charge density of glass which fills the moat. When the slope of the moat makes an acute angle with the pn junction plane in the more heavily doped one of the semiconductor regions for forming the pn junction (as conventionally referred to that the slope and the pn junction plane form a negative bevel), the blocking voltage of the pn junction can be improved by coating the moat with glass having negative surface charge in the case where an n-type semiconductor region is contiguous to the moat.
- A further investigation made by the present inventors has revealed that when glass having positive surface charge is provided in the moat of a semiconductor device having the above-mentioned negative bevel structure, it is easier to break down the pn junction in the bulk of the semiconductor substrate. It was also found that the field strength in the channel stopper region provided at the outer periphery of the moat is increased causing easier breakdown when a surface charge, even though negative, exceeds 9x10"/cm2, and that the negative surface charge is most preferably in a range from 3x10"/cm2 to 9x1011/cm2 for preventing easy breakdown of a pn junction and for reproducibly providing a high blocking voltage of the pn junction.
- It was also found that the characteristics of the glass-semiconductor interface is also influenced to some extent by the orientation of semiconductor surface and that the blocking voltage is little influenced by the thickness of the glass layer in case when the thickness is not less than about 15 pm.
- Furthermore, according to the present invention the depth of the channel stopper region from the major surface of the substrate is greater than the depth of the pn junction (J) from that major surface. This feature gives the device a high breakdown voltage, because when the pn junction is reverse biassed the depletion layer extends into the first one of the semiconductor regions. The excess depth of the channel stopper region prevents the concentration of electric field, and so prevents early breakdown, even when a high reverse bias is applied. The greater thickness of the channel stopper region also make the device easier to fabricate, as will be discussed later. In JP-A-54-14677, however, the thickness of the channel stopper region is less than the depth of the pn junction.
- Embodiments of the invention will now be described in detail, by way of example, with reference to the accompanying drawings, in which:
- Fig. 1A is a perspective view showing a chip of a gate turn off thyristor, being a first enbodiment of the present invention,
- Fig. 1B is a view taken along the line IB-IB' in Fig. 1A,
- Fig. 2 is a graph showing the relationship between the surface charge density NFB of the glass applied to the moat and the forward blocking voltage applied between the anode and cathode in the thyristor of Fig. 1,
- Fig. 3A is a perspective view showing a reverse blocking thyristor chip of corner gate type according to a second embodiment of the present invention,
- Fig. 3B is a view showing part of the cross section taken along the line lllB-lllB' in Fig. 3A.
- Fig. 4A is a perspective view showing a reverse blocking thyristor chip of centre gate type according to a third embodiment of the present invention,
- Fig. 4B is a view showing part of the cross section taken along a line intersecting the centre of one major surface in Fig. 4A,
- Fig. 5 is a three-component composition diagram for showing a vitrifiable range of ZnO-B203 Si02 glass applicable to the present invention,
- Fig. 6A and 6B are three-component composition diagrams showing the relationship between the composition of glass containing ZnO, B203 and SiO2 and the surface charge density NFB, wherein Fig. 6A shows the case where the firing temperature of glass is 630°C, and Fig. 6B shows the case where the firing temperature is 780°C.
- Figs. 7A and 7B are three-component composition diagrams showing the relationship between the composition of glass containing ZnO, B203 and Si02 and the coefficient of thermal expansion of the above glass, wherein Fig. 7A shows the case where the firing temperature of glass is 630°C, and Fig. 7B shows the case where the firing temperature is about 780°C.
- Figs. 1A and 1B show a gate turn off thyristor (hereinafter referred to as a "GTO") according to a first embodiment of the present invention. Referring to Figs. 1A and 1B, a
silicon substrate 1 has a pair ofmajor surfaces type base region 12, an n-type base region 11 and a p-type emitter region 15 is formed between themajor surfaces type base region 11 and the p-type base region 12 form the two semiconductor regions of the present invention. The n-emitter 13 is divided into a plurality ofstrip regions 13. Each of the n-emitter regions 13 thus formed is exposed at onemajor surface 101, and acathode electrode 31 is provided on the exposed surface of each n-emitter region. The p-base region 12 is exposed to themajor surface 101 so as to surround the n-emitter regions 13, and agate electrode 32 is provided on the exposed surface of the p-base region 12. - As shown in Fig. 1A, the
cathode electrode 31 includes a common connectingregion 310 for connecting in common divided electrode portions in ohmic contact with the respective n-emitter regions. The common connectingregion 310 is provided above the exposed surface of the p-base region 12, with an Si02 film (not shown) interposed between the common connectingregion 310 and the exposed surface of the p-base region 12 to insulate the common connectingregion 310 from thebase region 12. The p-emitter region 15 is provided in part of the othermajor surface 102 in approximate registry with the n-emitter regions. That is, the p-emitter region 15 is divided into a plurality of regions similar to and to oppose the divided n-emitter regions 13. A first n+-type anodeshort region 161 penetrates each of the p-emitter regions thus formed; at central part thereof. A second n+-type anodeshort region 162 is exposed at part of themajor surface 102 other than respective exposed surfaces of the p-emitter regions and respective exposed surfaces of the first n+-type anodeshort regions 161. Ananode electrode 33 is formed over all themajor surfaces 102. The first anodeshort regions 161 and the second anodeshort region 162 form a short circuit between the n-base region 11 and theanode electrode 33. Such a shorting structure is effective in improving the turn-off characteristic. - A
moat 51 in the form of a closed loop is cut in themajor surface 101 along the periphery thereof. Themoat 51 is cut to a depth such that the bottom of the moat is within the high resistivity n-base region 11. Therefore, the pn junction formed between the p-base region 12 and the high resistivity n-base region 11 terminates at an inclined surface of themoat 51. An nt-typechannel stopper region 14, which has a greater thickness than the p-base region 12, is formed as a closed loop outside themoat 51.Glass material 21 having characteristics which will be explained later, is fired in themoat 51. Part of themajor surface 101 other than the part provided with the electrodes is coated with an Si02 film 41 for the purpose of passivation. The field mitigating means in the present embodiment are made up of theglass material 21 and the nT-typechannel stopper region 14 which has a greater thickness than the p-base region 12 and is highly doped. - Such a GTO can be fabricated, for example, by impurity diffusion techniques, photolithography and selective etching techniques, using an n-type (111) silicon plate having a resistivity of about 6-Qcm as the starting material. Gallium is diffused into the high resistivity n-type silicon plate from a pair of major surfaces thereof to form the P-
base region 12 and p-emitter region 15. Next, phosphorus is selectively diffused from the major surfaces into those areas corresponding toregions channel stopper region 14 is formed on the side of themajor surface 101, and the first anodeshort regions 161 and second anodeshort region 162 are formed on the side of themajor surface 102. Then, phosphorus is diffused into those areas corresponding to the n-emitter regions 13 so that the diffusion depth of the phosphorus is smaller than the depth of the p-base region 12, to form the n-emitter regions 13. - The
moat 51 can be made using selective etching techniques, and theelectrodes - Each n-
emitter region 13 has a thickness of about 15 µm, and the surface of diffusion has an impurity concentration of about 1020 atoms/cm3 or more. The p-base region 12 has a thickness of about 30 µm, and the surface of diffusion has an impurity concentration of about 1018 atoms/cm3. The n+-typechannel stopper region 14 has a thickness of about 70 µm and a width of about 50 µm or more, and the surface of diffusion has an impurity concentration of about 1020 atoms/cm3 or more. The n-base region 11 has a thickness of about 180 pm and a resistivity of about 60 Ocm. Themoat 51 has a width of about 300 µm and a depth of about 90 um, and therefore the depth of the moat is greater than the thickness of thechannel stopper region 14 and also than that of the p-base region 12. - The
glass 21 is filled in themoat 51 provided in the above-mentioned GTO. Theglass 21 can be formed by suspending glass powder in a liquid, adhering the powder to the surface of themoat 51 by electrophoresis, and then firing the powder. Theglass 21 obtained after firing has a thickness of 20 to 30 µm. Theglass material 21 used was zinc borosilicate glass. The zinc borosilicate glass preferably contains more than about 50 percent ZnO by weight. - The amount of surface charge on the
glass 21 can be varied by changing the composition of theglass 21, the firing temperature, firing time, firing atmosphere and the cooling rate after firing of theglass 21. Changes in the forward blocking voltage of the above-mentioned GTO (namely, changes in the blocking voltage of the pn junction formed between the p-base region 12 and the n-base region 11) were measured when the amount of charge kept by theglass 21 was varied. The results of measurements are shown in Fig. 2. The amount of charge per unit area at the surface of theglass 21 which is in contact with the inner wall of themoat 51, is expressed by NFB (cm-2). - As is evident from Fig. 2, the blocking voltage is highest when the surface charge density NFB is within a range from -3x1011/cm2 to -9x10"/ cm2, that is, when the negative surface charge per unit area lies within a range from 3x1011/cm2 to 9x10"/cm2. Even when the surface charge density NFB varies to some extent within this range, the high blocking voltage is not affected. In other words, a reproducible, high blocking voltage is obtained, even if the surface charge density NFB fluctuates. Accordingly, in the present embodiment, it is preferred that the
glass 21 has a surface charge density NFB of -3x 1011/cm2 to -9x1011/cm2. - According to experiments made by the present inventors, when zinc borosilicate glass was fired at a temperature of 680 to 740°C for about 40 minutes in an atomspere containing 1 to 100 weight percent dry 02 (and the balance N2), and then cooled at a rate of about 1°C/min., a value of NFB within the above-mentioned range was obtained. The zinc borosilicate glass had, for example, 65 percent ZnO, 21 percent B203, 8.5 percent Si02, 4 percent PbO, 1 percent SnO and 0.5 percent Sb203 by weight.
- In the present invention, the depth of the n *-type
channel stopper region 14 is larger than that of the p-base region 12. Such a structure has the following advantages. Firstly, it is not necessary that a p-type impurity is selectively diffused into thesilicon substrate 1 from themajor surface 101 to form the p-base region 12. That is, when the p-type impurity is diffused to a predetermined depth from the whole of themajor surface 101 and then an n-type impurity is selectively diffused into thechannel stopper region 14 in such a manner that the concentration of the n-type impurity is higher than that of the p-type impurity and the diffusion depth of the n-type impurity is greater than that of the p-type impurity, the channel stopper region shown in Fig. 1B is formed. Therefore, the fabricating process can be simplified. Such a fabricating method is particularly advantageous when gallium and aluminium are diffused as the p-type impurity, since it is difficult to obtain an appropriate mask for selective diffusion of gallium and aluminium. Also, suppose that theregion 14 were smaller in thickness than theregion 12. In this case, it would be required to selectively form theregion 12 in that part of themajor surface 101 which is surrounded by themoat 51. If, in the above selective formation process, pin holes are accidentally generated in a cover for masking a portion of themajor surface 101 outside themoat 51 from a diffusion source, p-type regions which are greater in depth than theregion 14, are formed in the surface portion outside themoat 51, and the function of the channel stopper region is deteriorated. In other words, the effective width of the channel stopper region is reduced. Further, the yield of high breakdown voltage semiconductor devices is lowered. According to the present embodiment, all of these drawbacks may be eliminated. - Secondly, the gradient of impurity concentration at the boundary between the n+-type
channel stopper region 14 and the n-base region 11 becomes gentle, since the thickness of theregion 14 is large. Therefore, when the pn junction exposed at themoat 51 is reversely biased, the depletion layer extends from the pn junction into the n-base region 11 and, even after it has reached thechannel stopper region 14, it can be further extended gradually in thechannel stopper region 14 when the reverse bias voltage is increased. Thus, even when a high reverse bias voltage is applied, concentration of the electric field is mitigated, and therefore it is possible to provide a high breakdown voltage. - The width of the channel stopper region 14 (that is, the width of the
region 14 at the boundary between theregion 14 and the n-base region 11 in the direction perpendicular to the lengthwise direction of the moat 51) is preferably made larger than 50 pm, in order effectively to extend the depletion layer and to make the breakdown voltage high. - The depth of the
moat 51 is preferably larger than respective thicknesses of the p-base region 12 andchannel stopper region 14. When the depth of themoat 51 is smaller than the thickness of the p-base region 12 so that part of the p-base region 12 is extended along the bottom of themoat 51 and the pn junction J terminates at the bottom of themoat 51, the breakdown voltage varies widely. This is because the electric field generated at one end of the pn junction J, when the pn junction J is reverse biased, depends upon the length of that portion of the p-base region which is located beneath the bottom of themoat 51 and upon the depth of themoat 51 in a critical manner, and cannot be easily controlled. - Also, the difference in depth between the bottom of the
moat 51 and the bottom of the region 12 (namely, the pn junction J) is preferably made larger than 10 pm to reduce variations in the breakdown voltage. The reason for this is as follows. Compensation of impurity concentration takes place in the proximity of a pn junction, since two regions of different conductivity types are contiguous to each other at the pn junction, so that compensated regions have higher resistivity. Thus, the depletion layer can be readily extended in such regions. When the pn junction is located in the proximity of the bottom of themoat 51, the depletion layer can be easily extended to reach the bottom of themoat 51 and further extension of the depletion layer is restricted. Therefore, the blocking voltage can be lowered at a portion beneath themoat 51, and the leakage current can be increased. - When the channel stopper region is formed by diffusion, it is desirable to make time required for the formation of the channel stopper region as short as possible. Accordingly, it is preferable to make the thickness of the channel stopper region smaller than the depth of the
moat 51. - Now, a second embodiment of the present invention will be explained with reference to Figs. 3A and 3B. The second embodiment relates to a reverse blocking thyristor. Unlike the above-mentioned GTO having a shorting structure in the anode region, the reverse blocking thyristor must be able to block both forward and reverse voltages applied between anode and cathode electodes. That is, a high blocking voltage is desired in both forward and reverse directions.
- In Figs. 3A and 3B, parts similar or equivalent to those' in Figs. 1A and 1B are given the same reference numerals as in Figs. 1A and 1B.
- Referring to Figs. 3A and 3B, two
moats major surface 101, and the depth and width of each of themoats moat 51 shown in Fig. 1B. The distance between themoats gate electrodes major surface 101 which is surrounded by theinner moat 511. Thegate electrode 32 is near one corner of the above-mentioned inner part, and has substantially a rectangular form. Thecathode electrode 31 is insulated from thegate electrode 32, and is formed on the rest of the above inner part. Such an electrode arrangement is called a corner gate type. - Fig. 38 shows part of a cross section taken along the line IIIB―IIIB' in Fig. 3A. Referring to Fig. 3B, an n-
emitter region 13, a p-base region 12, an n-base region 11 and a p-emitter region 15 are stacked in the order described, between thecathode electrode 31 and theanode electrode 32 formed on the othermajor surface 102. The n-base 11 region and the p-base region 12 form the two semiconductor regions of the present invention. Unlike the embodiment shown in Figs. 1A and 1B, the n-emitter region 13 is not divided into a plurality of regions, and the p-emitter region 15 has no shorting structure and is uniform in structure. A pn junction J, between the n-base region 11 and the p-emitter region 15 terminates at theouter moat 512 due to the presence of a p+-type region 151 (communicating with the p-emitter region 15) which connects peripheral parts of themajor surfaces base region 12 and the n-base region 11 terminates at the inner inclined surface of theinner moat 511, and the pn junction J2 and the inner inclined surface of themoat 511 form a negative bevel structure. An n+-typechannel stopper region 14 is formed between themoats major surface 101, and is in the form of a closed loop.Glass material 21 having negative charge is filled in each of themoats major surface 101 in the area other than the gate and cathode electrode areas. In the second embodiment, field mitigating means include themoats glass material 21 and thechannel stopper region 14. - The semiconductor element shown in Figs. 3A and 3B can be fabricated by e.g. impurity diffusion techniques, photolithography, and selective etching techniques, as for the embodiment shown in Figs. 1A and 1B. Further, the same materials as used in the embodiment shown in Fig. 1A and 1B can be used. The thickness of the
region 14 is greater than that of the n-emitter region 13 as in the embodiment shown in Figs. 1A and 1B. The p+-type region 151 may be formed e.g. by diffusing a p-type impurity from respective peripheral portions of themajor surfaces - In the forward blocking state of the second embodiment, namely, when an input voltage is applied between the anode and
cathode electrodes anode electrode 33 is positive compared with that of thecathode electrode 31, a depletion layer extends from the pn junction J2 mainly into the n-base region 11. Since theglass material 21 has negative charge, the depletion layer is more easily extended along themoat 511 rather than into the bulk of the semiconductor substrate in the area away from the moat. Therefore, electric field in the region around the moat is mitigated. When the voltage applied between the anode and cathode electrodes is increased, the end of the depletion layer will reach thechannel stopper region 14. Since theregion 14 has a higher impurity concentration than the n-base region 11, the rate of extension of the depletion layer will be largely slowed down in thechannel stopper region 14. Thus, the depletion layer is prevented from being further extended outward and from being exposed to the periphery of the semiconductor substrate. Accordingly, it is possible to maintain a high blocking voltage. The rate of extension of the depletion layer in the portion of theregion 11 along the moat is determined mainly by the amount of charge kept by theglass material 21. Experiments made by the present inventors have shown that a surface charge density NFB of theglass material 21 is preferably in a range from -3x 10"cm-2 to -9x1011cm-2to provide a breakdown voltage of about 1.6 kV, which has not been obtained heretofore. - When the polarity of the voltage applied between the anode and cathode electrodes is reversed with the potential of the cathode electode being positive as compared with that of the anode electrode, the thyristor takes a reverse blocking state. At this time, a depletion layer extends from the pn junction J, mainly into the n-
base region 11. In the part along themoat 512, the rate of extension of the depletion layer becomes slow in thechannel stopper region 14 in accordance with the same mechanism as in the forward blocking state. Accordingly even when a higher reverse voltage is applied, the depletion layer is not extended along themoat 511 and does not reach the p-base region 12, and the blocking state is not destroyed. - It is to be noted in the second embodiment that the pn junction J1 and the outer inclined surface of the
moat 512 form a positive bevel structure, whereas the pn junction J2 and the inner inclined surface of themoat 511 form a negative bevel structure. - In such a situation, if the channel stopper region is absent and the
glass 21 has negative charge, the blocking ability in the forward blocking state is good but the blocking ability in the reverse blocking state is not so good. On the other hand, if the channel stopper region is present and theglass 21 has positive charge, the blocking ability in the reverse blocking state is good but the blocking ability in the forward blocking state is not so good. In the second embodiment, the blocking function is good in both the forward and reverse blocking states, since the channel stopper region is formed and theglass 21 has negative charge. - Next, a third embodiment of the present invention will be explained with reference to Figs. 4A and 4B. In the third embodiment, a
semiconductor substrate 1 has a disc shape. Eachmajor surface major surface 101 is smaller than that of themajor surface 102. Theside surface 103 connecting themajor surfaces major surfaces major surfaces gate electrode 32 is formed on a central part of themajor surface 101, and acathode electrode 31 is formed on themajor surface 101 so that it is insulated from and surrounds thegate electrode 32. Amoat 51 in the form of a closed loop is formed around thecathode electrode 31. Ananode electrode 33 is formed on the whole of the othermajor surface 102. The above-mentioned third embodiment is a reverse blocking thyristor having the same function as the reverse blocking thyristor shown in Figs. 3A and 3B. The thyristor of this embodiment as shown in Fig. 4A is called the reverse blocking thyristor of centre gate type, since the gate elctrode is arranged at a central portion of one major surface. - Fig. 4B shows a part of a cross section taken along a diameter of the circular plate in Fig. 4A. Referring to Fig. 4B, an n-
emitter region 13, a p-base region 12, an n-base region 11 and a p-emitter region 15 are stacked beween thecathode electrode 31 and theanode electrode 33, as in Fig. 3B. The n-base region 11 and the p-base region 12 form the two semiconductor regions of the present invention. However, unlike the second embodiment shown in Fig. 3B, the p-emitter region 15 is uniform and terminates at theside surface 103 of thesemiconductor substrate 1. The p-emitter region 15 has a higher impurity concentration than the n-base region 11. Theside surface 103 is tapered toward themajor surface 101 such that the area of a plane parallel to themajor surfaces major surface 102 to 101. Therefore, the pn junction J1 which is formed between the p-emitter region 15 and the n-base region 11, and theside surface 103 form a positive bevel structure. Such a positive bevel structure acts to increase the blocking voltage of the pn junction J" as described in US―A―3491272. The pn junction J2 between the p-base region 12 and the n-base region 11 terminates at the inner, inclined surface of themoat 51. The p-base region 12 has a higher impurity concentration than the n-base region 11. Therefore, the inner inclined surface of themoat 51 and the pn junction J2 form - a negative bevel structure. The depth of themoat 51 is greater than the depth of the pn junction J2. A pn junction J3 between the n-emitter region 13 and the p-base region 12, and the pn junction J2 terminate at the inner inclined surface of themoat 51. There is an n+-typechannel stopper region 14 between the outer inclined surface of themoat 51 and theside surface 103 to connect these surfaces with each other. Thechannel stopper region 14 is thicker than the n-emitter region 13.Glass material 21 which has an appropriate composition is applied in themoat 51 and fired to have negative charge. - In the semiconductor element of the third embodiment, field mitigating means include the
moat 51, theglass 21 in the moat and the positive bevel structure formed by the pn junction J, and theside surface 103. - The semiconductor element shown in Figs. 4A and 4B can be fabricated in similar manner as the second embodiment shown in Figs. 3A and 3B, and therefore a detailed explanation is omitted. Further, the same materials as used in the second embodiment can be used in the third embodiment. When the thyristor of this embodiment shown in Figs. 4A and 4B takes a forward blocking state, a high blocking voltage is obtained on the basis of the same mechanism as in the embodiments shown in Figs. 1A and 3A. On the other hand, when the thyristor has a reverse blocking state (i.e., when the greater part of a voltage applied across the thyristor is applied across the pn junction J,), a high breakdown voltage is obtained on the basis of the positive bevel structure at the
side surface 103. - Next, explanation will be made on glass which can be used as the
glass material 21 in the above-mentioned embodiments. This glass essentially consists of three main components ZnO, B203 and Si02. Similar glass material is disclosed in DE-A-2851479. - Various materials containing three components ZnO, B203 and Si02 at different mixing ratios were prepared to find a vitrifiable composition range. The results obtained are shown in Fig. 5. In Fig. 5, a region A indicates a composition range in which transparent glass was formed, and a region B indicates a composition range in which any mixed material was devitrified or divided into a plurality of phases and therefore transparent glass was not obtained.
- Next, a glass material having a composition in the vitrifiable range shown in Fig. 5A was fused in a platinum crucible, and then quenched in the pure water to obtain a lump of glass, which was finely pulverized by a ball mill to obtain a glass powder for passivation. Various glass powders having different compositions within the vitrifiable range were prepared. Each of these glass powder materials was fired on a silicon substrate at a temperature in a range from 630°C to 780°C, and the surface charge density NFB was measured. In such a temperature range, glass is readily softened and is not crystallised. Fig. 6A shows equi-NFB curves which were obtained when the glass powder materials were fired at 630°C, Fig. 6B shows equi-NFB curves for a firing temperature of 780°C. As is apparent from Figs. 6A and 6B, the value of NFB varies greatly with the composition, namely, the mixing ratio of the three components, and moves to the negative side as the firing temperature is higher.
- A composition range of glass and a range of firing temperature which are preferably applied to the present invention, are known from Figs. 6A and 6B.
- When glass is applied on a semiconductor material, it is desirable that the coefficient of thermal expansion of the glass is as close as possible to that of the semiconductor material. When a large number of elements or chips are simultaneously formed in a semiconductor wafer having a large diameter and then are divided into the respective elements, moats as described in the above-mentioned embodiments are distributed almost all over one major surface of the wafer with glass filled therein. In such a case, if the above glass has a greatly different thermal expansion coefficient from that of the semiconductor wafer, the semiconductor wafer may be warped due to the different thermal expansion (contraction) during cooling from the firing temperature to the room temperature. When a photo-etching process is to be carried out for the wafer (for example, when the gate and cathode electrodes are to be formed through the photo-etching process in the above-mentioned embodiments), such a warp may constitute an obstacle to the accurate printing of a predetermined photoresist pattern on the wafer. Further, when the warp of the wafer is large, there is a risk of the wafer and glass being cleaved and/or cracked. In the case of silicon, to avoid such difficulties, it is preferable to set the coefficient of thermal expansion of glass after firing within a range from 25x10-7/°C to 50xlO-7PC.
- The glass containing three main components ZnO, B203 and Si02 as described above has the thermal expansion characteristics discussed below. Several glass powder materials having various compositions within the vitrifiable range shown in Fig. 5 were each fired to a thickness of 10 pm on a silicon wafer of 200 to 500 pm thick at a temperature within a range from 630 to 780°C, and the relation between the warp of silicon wafer and the difference in coefficient of thermal expansion between glass and silicon was studied. As a result of the study, it was found that the coefficient of thermal expansion of glass varied greatly with the composition of glass and the firing temperature. Fig. 7A shows equi-thermal-expansion curves in the case where glass powder materials were fired at 630°C, and Fig. 7B shows equi-thermal-expansion curves for a firing temperature of 780°C. As is apparent from the comparison between Figs. 7A and 7B, glass containing a large amount of ZnO has a smaller coefficient of thermal expansion, and the equi-thermal-expansion curves move to a composition region containing a relatively smaller amount of ZnO, as the firing temperature is higher. When glass had a coefficient of thermal expansion of 25x10-7/°C to 50x10-7/°C after firing, the warp of silicon wafer was small and it was possible to print precisely a photoresist pattern for photo-etching. Moreover, the glass was not cracked. On the other hand, when glass had a coefficient of thermal expansion less than 25x10-7/°C or more than 50x10-7/°C after firing, the glass might be broken, or the warp of silicon wafer became large and therefore it was difficult to print the photoresist pattern.
- A composition range of glass and a range of firing temperature, which are preferable in the present invention, may be determined from Figs. 7A and 7B. Since the above-mentioned glass contains only three main components ZnO, B203 and Si02, variations in characteristic of glass are limited to a relatively small range. Therefore, when the above glass is used, uniform, reproducible semiconductor elements may be obtained.
- While the preferred embodiments have been explained, the present invention is not limited to these embodiments but is applicable to various semiconductor devices. For example, when the p-
emitter region 15 in the embodiment shown in Figs. 1A and 1B is replaced by an n-type semiconductor region, a transistor structure is obtained. Such a structure is also within the scope of the present invention. Also, in the embodiment shown in Fig. 4B, theside surface 103 of thesemiconductor substrate 1 may be coated with a passivation material such as glass or resin. The materials, fabricating methods and dimensions used in the embodiments may be appropriately altered. The conductivity type of each semiconductor region may be reversed. In this case, glass having positive charges is employed, and a preferable range of surface charge density is +3x10"cm-2 to +9x1011cm-2.
Claims (7)
characterised in that:
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE8181306038T DE3174437D1 (en) | 1981-12-22 | 1981-12-22 | Semiconductor device passivated with glass material |
EP19810306038 EP0082224B1 (en) | 1981-12-22 | 1981-12-22 | Semiconductor device passivated with glass material |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP19810306038 EP0082224B1 (en) | 1981-12-22 | 1981-12-22 | Semiconductor device passivated with glass material |
Publications (2)
Publication Number | Publication Date |
---|---|
EP0082224A1 EP0082224A1 (en) | 1983-06-29 |
EP0082224B1 true EP0082224B1 (en) | 1986-04-16 |
Family
ID=8188484
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP19810306038 Expired EP0082224B1 (en) | 1981-12-22 | 1981-12-22 | Semiconductor device passivated with glass material |
Country Status (2)
Country | Link |
---|---|
EP (1) | EP0082224B1 (en) |
DE (1) | DE3174437D1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE4109533C2 (en) * | 1990-09-13 | 1994-06-23 | Eupec Gmbh & Co Kg | Passivated semiconductor component and method for its production |
DE102016124670B4 (en) * | 2016-12-16 | 2020-01-23 | Semikron Elektronik Gmbh & Co. Kg | Thyristor with a semiconductor body |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5474817A (en) * | 1977-11-28 | 1979-06-15 | Hitachi Ltd | Glass for coating semiconductor device |
GB2047461A (en) * | 1979-04-19 | 1980-11-26 | Philips Electronic Associated | Semiconductor device |
-
1981
- 1981-12-22 DE DE8181306038T patent/DE3174437D1/en not_active Expired
- 1981-12-22 EP EP19810306038 patent/EP0082224B1/en not_active Expired
Also Published As
Publication number | Publication date |
---|---|
DE3174437D1 (en) | 1986-05-22 |
EP0082224A1 (en) | 1983-06-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US5804868A (en) | Semiconductor device having planar junction | |
US5086332A (en) | Planar semiconductor device having high breakdown voltage | |
JP3292651B2 (en) | Photovoltaic device | |
US4484214A (en) | pn Junction device with glass moats and a channel stopper region of greater depth than the base pn junction depth | |
US3936857A (en) | Insulated gate field effect transistor having high transconductance | |
EP1394860B1 (en) | Power devices with improved breakdown voltages | |
GB1219986A (en) | Improvements in or relating to the production of semiconductor bodies | |
EP0018730B1 (en) | Semiconductor device having a high breakdown voltage | |
JPH0358185B2 (en) | ||
US5767548A (en) | Semiconductor component with embedded fixed charges to provide increased high breakdown voltage | |
US4275408A (en) | Thyristor | |
JP2870553B2 (en) | High voltage semiconductor device | |
US4730208A (en) | Semiconductor device | |
JPH11330456A (en) | Semiconductor device | |
US4420765A (en) | Multi-layer passivant system | |
US3559006A (en) | Semiconductor device with an inclined inwardly extending groove | |
EP0082224B1 (en) | Semiconductor device passivated with glass material | |
US4651178A (en) | Dual inverse zener diode with buried junctions | |
EP0178664A2 (en) | Solid state image sensing device and method for making the same | |
US4125415A (en) | Method of making high voltage semiconductor structure | |
KR100326487B1 (en) | Support Lead Frame for Supporting Semiconductor Chip | |
GB2069235A (en) | Protecting semiconductor devices | |
US4942446A (en) | Semiconductor device for switching, and the manufacturing method therefor | |
WO1994028586A1 (en) | Semiconductor device having high breakdown strength | |
US3462657A (en) | Protection means for surface semiconductor devices having thin oxide films therein |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 19830304 |
|
AK | Designated contracting states |
Designated state(s): DE FR GB NL |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Kind code of ref document: B1 Designated state(s): DE FR GB NL |
|
REF | Corresponds to: |
Ref document number: 3174437 Country of ref document: DE Date of ref document: 19860522 |
|
ET | Fr: translation filed | ||
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
26N | No opposition filed | ||
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: NL Payment date: 19871231 Year of fee payment: 7 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: DE Payment date: 19881227 Year of fee payment: 8 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: GB Effective date: 19891222 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: NL Effective date: 19900701 |
|
NLV4 | Nl: lapsed or anulled due to non-payment of the annual fee | ||
GBPC | Gb: european patent ceased through non-payment of renewal fee | ||
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: FR Effective date: 19900831 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Effective date: 19900901 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: ST |